Abstract
The application of intelligent algorithms in the treatment of intractable pain of patients with platelet-rich plasma (PRP) knee osteoarthritis by magnetic resonance was investigated. The automatic diagnosis of magnetic resonance knee osteoarthritis was established with multiple intelligent algorithms, including gray projection algorithm, adaptive binarization algorithm, and active shape model (ASM). The difference between automatic magnetic resonance detection indexes of the patients with knee osteoarthritis and artificial measurement results was analyzed. The included patients received PRP treatment. Knee osteoarthritis MRI osteoarthritis knee scores (KOA MOAKS) and Western Ontario and McMaster Universities arthritis index (WOMAC) before and after treatment were compared. The results showed that the results of knee osteoarthritis scores, inferior angle of femur, superior angle of tibia, and tibiofemoral angle (TFA) by automatic magnetic resonance diagnostic model were entirely consistent with artificial detection results. After the treatment, the total scores of knee lateral area, interior area, central area, and patellar area were all remarkably lower than those before the treatment (P < 0.05). After the treatment, knee KOA MOAKS scores and WOMAC scores were both lower than those before the treatment (P < 0.05). Visual analogue scale (VAS) scores 1 week, 2 weeks, and 3 weeks after the treatment were decreased compared with those before the treatment (P < 0.05). Relevant studies indicated that intelligent algorithm-based automatic magnetic resonance diagnostic knee osteoarthritis model showed good utilization values, which could provide the reference and basis for the treatment of the patients with knee osteoarthritis.